A method is proposed to increase the scientific usefulness of a mathematical model of respiratory physiology as an encoder of the scientific literature. The method regards the simulation product to be a generalized n-tuple for a relational data base and specific laboratory experiments or reports as views (m-tuples) on that data base. The model thus becomes a data base generator and all relations are implicit in the source algorithms. By using both control flow and data flow analysis on the source code, statements about mechanism, altered simulations, fact retrieval, and hypothesis test should be rendered automatic. Query handling for the subject knowledgeable scientist is the principal objective, along with the condensed encoding of respiratory physiology, while most of the research addresses the problem of creating a single working assemblage, new problems arise because (1) the m-tuples must be ordered by the query, and (2) many physiological concepts represent patterns on the columns of a relational data base.